As we embark on our second season of data collection in Lithuania, we highlight what we learned from analysing our first season’s data.
While raw figures might make interesting reading, when it comes to thinking about bycatch, we need to analyse our data to draw useful conclusions about how effective our net panels might be in reducing seabird mortality in gillnet fisheries.
One of the difficulties with almost all bycatch data, irrespective of fishery or accidentally captured species, is that it is usually a rare, sporadic event – this means that our data set contains lots of zeros, when there was no bycatch at all. Most means of analysis don’t cope particularly well with this – but thankfully, there are ways!
Without getting into the details (we will aim to publish our results in full once we have completed our next field season), the mean number of birds caught per set was about a third lower in our experimental sets with net panels (mean of 0.37 birds/set) than control sets (0.54/set). While this is an encouraging result, this difference was not statistically significant. Given this outcome, we are undertaking further trials this winter in Lithuania, continuing to conduct paired trials to bottom-out whether these net panels might have utility in reducing seabird bycatch in the Baltic. The signs are encouraging, but more work is required!
One interesting (and statistically significant!) result is that total fish catch was actually higher in the nets carrying panels compared to standard gillnets (a graph below if you’re interested!). This is an important result, as technical mitigation is unlikely to be popular with fishermen if it negatively affects catch. We’ll continue to keep a close eye on how net panels affect target catch…

Total fish catch in control and experimental nets. Horizontal lines are the median, boxes the interquartile range, error bars the 95% CI, and dots are final outliers
Thanks to Alex Bond of RSPB for carrying out the analysis!